New opportunity

Chemistry Content Expert – Agentic Solutions

RELX · India

About this role

The Chemistry Content Expert – Agentic Solutions will bridge the gap between deep chemistry expertise and Generative AI search technologies. This role focuses on optimizing how researchers discover and interact with chemistry data within the Reaxys platform by designing and validating AI-driven search capabilities. Responsibilities: Support the development and validation of chemistry connectors for agentic solutions. Create high-quality test sets and evaluation datasets to assess the accuracy, quality, and reliability of these connectors. Analyze connector outputs against expected results to identify gaps, errors, or edge cases. Perform data analysis and generate reports summarizing evaluation findings and recommendations. Collaborate with cross-functional Data Science and Technology partners to support issue resolution and shape new AI search solutions. Requirements: PhD or post-doctoral experience in Organic Chemistry or a closely related field, with hands-on laboratory or research experience. Ability to design evaluation/test datasets and assess the quality of outputs from chemistry-related tools. Understanding of how chemistry data is structured, represented, and retrieved in scientific information systems. Strong analytical and communication skills. Familiarity with or interest in AI/NLP-based or agentic systems is considered a plus. Benefits: The company offers a comprehensive benefits package including health insurance for the employee and family, group life and accident insurance, flexible working arrangements, an employee assistance program, modern family benefits, and various paid time off. Subsidized meals and transport services are available for employees based in Chennai.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Organic ChemistryGenAIData EvaluationTest Set CreationData AnalysisScientific Information SystemsNLPAgentic Systems

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Design and Validation of Evaluation Datasets
The core of this role is creating high-quality test sets to assess the accuracy and reliability of AI-driven chemistry connectors. You must be able to translate complex chemical queries into structured evaluation benchmarks that effectively stress-test the system's search capabilities.
Cross-Functional Collaboration with Data Science
You will act as the bridge between deep chemistry domain knowledge and technical AI development. Success depends on your ability to communicate findings, explain chemical edge cases to non-chemist engineers, and work iteratively to resolve performance issues in agentic search solutions.
Analysis of AI Search Performance
Beyond creating tests, you are responsible for analyzing the outputs of chemistry connectors against expected results. This requires identifying patterns in errors, documenting edge cases, and generating actionable reports that guide the technical team in refining the AI's search behavior.

Likely challenges to think through

Bridging Domain Expertise and Technical Implementation
A significant challenge is effectively communicating the nuances of organic chemistry to technical partners who may lack a scientific background. Prepare by practicing how to explain chemical data structures and search requirements in a way that is actionable for developers building agentic systems.
Managing Ambiguity in Agentic Search Evaluation
Agentic systems can produce non-deterministic results, making it difficult to define a single 'correct' answer for every query. Prepare by considering how to design evaluation frameworks that account for scientific accuracy while allowing for the flexibility required in generative search.

Practice your answers

Technical practice

Designing Evaluation Datasets for Chemistry AI

What to coverExplain your methodology for creating high-quality test sets that accurately measure the performance of AI-driven chemistry search tools. Focus on how you ensure the dataset covers a representative range of chemical structures, reaction types, and potential edge cases.

Validating Chemistry Connectors

What to coverDescribe the process you would use to validate a chemistry connector within an agentic system. Discuss how you define 'expected results' and the metrics you would use to quantify accuracy, reliability, and precision.

Handling Ambiguity in Chemical Information Retrieval

What to coverChemistry data often contains synonyms, varying nomenclature, and complex structural representations. Discuss how you would evaluate an AI system's ability to correctly interpret and retrieve information despite these inherent complexities.

Analyzing AI Output Errors

What to coverWhen an AI connector produces an incorrect or unexpected result, what steps do you take to diagnose the root cause? Explain how you would categorize errors (e.g., hallucination, retrieval failure, or logic error) and report these findings to technical partners.

Structuring Chemistry Data for Search

What to coverHow does your understanding of chemical taxonomies and database structures inform your approach to improving search retrieval? Suggest how you might bridge the gap between raw chemical data and the requirements of a generative AI model.

Agentic Systems vs. Traditional Search

What to coverWhat are the unique challenges in evaluating 'agentic' search solutions compared to traditional keyword-based search? Address the dynamic nature of agentic workflows and how you would maintain evaluation consistency.

Reporting Evaluation Findings

What to coverDescribe your approach to summarizing complex technical evaluation results for non-specialist stakeholders. Focus on how you translate data analysis into actionable recommendations for product improvement.

Integrating Domain Expertise with NLP

What to coverHow do you leverage your background in organic chemistry to improve the performance of NLP models? Provide an example of how domain-specific knowledge can be used to refine training data or prompt engineering.

Behavioural practice

Cross-functional Collaboration

What to coverDescribe a time you worked with technical teams (such as Data Science or Engineering) to solve a complex problem. How did you ensure your chemistry expertise was clearly understood and integrated into the technical solution?

Managing Conflicting Priorities

What to coverTell me about a situation where you had to balance the need for high-quality, rigorous data evaluation with tight project deadlines. How did you prioritize your tasks to ensure the most critical issues were addressed first?

Adapting to New Technologies

What to coverGive an example of how you have learned or applied a new technology or methodology in your research. How did you overcome the initial learning curve to become effective in your role?

Communicating Technical Recommendations

What to coverDescribe a time when you identified a significant gap or error in a system's output. How did you communicate this to your team, and what was the result of your intervention?

Sharpen your knowledge

Topics to prepare

Agentic AI and Search Architectures
Review the conceptual framework of agentic systems, specifically how they differ from traditional RAG (Retrieval-Augmented Generation) in chemistry search. Exercise: Research how autonomous agents decompose complex scientific queries into multi-step retrieval tasks. This is relevant for designing and validating the chemistry connectors mentioned in the job description.
Scientific Data Representation
Study how organic chemistry data is structured in databases, including SMILES strings, InChI keys, and chemical taxonomies. Exercise: Map a set of complex organic reactions to a structured format suitable for machine learning ingestion. This is essential for understanding how to optimize data retrieval within the Reaxys platform.

Skills to revise

Evaluation Dataset Design
Develop the ability to create 'Gold Standard' test sets for AI model validation. Exercise: Create a test set of 20 chemistry-specific queries with expected, verified outputs, including edge cases like ambiguous nomenclature or complex reaction conditions. This directly addresses the requirement to assess the accuracy and reliability of AI connectors.
Error Analysis and Reporting
Practice systematic identification of AI output failures. Exercise: Take a set of AI-generated search results and categorize them by error type (e.g., hallucination, retrieval failure, or incorrect chemical interpretation). Draft a brief report summarizing the findings and suggesting improvements. This is a core responsibility for shaping new search solutions.

Your preparation checklist

Develop a framework for evaluating AI-driven chemistry search
Prepare a document outlining how you would design a test set to validate the accuracy of a chemistry-focused AI agent. Include your methodology for defining 'ground truth' in chemical data retrieval and how you would categorize errors such as hallucinations, structural misinterpretations, or irrelevant search results.
Create a sample evaluation report
Draft a mock report summarizing findings from a hypothetical evaluation of a chemistry search tool. Demonstrate your ability to analyze connector outputs, identify specific edge cases in chemical nomenclature or reaction data, and provide actionable recommendations for technical teams to improve system performance.
Map your research experience to data-centric problem solving
Prepare specific examples from your PhD or post-doctoral research that demonstrate your ability to handle large, complex datasets. Focus on instances where you had to structure, clean, or validate scientific information, as this directly mirrors the requirements for managing chemistry data in information systems.
Prepare questions regarding agentic workflows
Formulate thoughtful questions for the interviewers about how the team handles the 'agentic' aspect of the role, such as how they manage multi-step reasoning in chemistry queries or how they balance automated retrieval with human-in-the-loop validation processes.
Review Reaxys and scientific information system architecture
Research the general structure of scientific databases and taxonomies. Be prepared to discuss how you would approach the challenge of ensuring that AI models correctly interpret chemical structures, reaction conditions, and bibliographic data when retrieving information for researchers.

Availability last checked: 18 Oct 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…